NVIDIA H100 SXM5 80GB meets the minimum VRAM requirement for Q4 inference of trl-internal-testing/tiny-random-LlamaForCausalLM. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
NVIDIA H100 SXM5 80GB can run trl-internal-testing/tiny-random-LlamaForCausalLM with Q4 quantization. At approximately 471 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 76GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 4GB | 80GB | 470.65 tok/s | ✅ Fits comfortably |
| Q8 | 7GB | 80GB | 352.82 tok/s | ✅ Fits comfortably |
| FP16 | 15GB | 80GB | 195.83 tok/s | ✅ Fits comfortably |